On 7 April 2026, Montreal-based company Winston AI launched a forensic image detection service designed to identify deepfakes, synthetic media, and manipulated visual content. Rather than offering a basic binary verdict, the tool pinpoints specific modified areas down to the pixel and identifies the exact artificial intelligence model or editing software used to alter the image. The platform operates a forensic pipeline combining six independent analytical techniques—such as analysing compression artifacts, camera sensor fingerprints, and noise anomalies—with cryptographic provenance analysis. These forensic signals are synthesised by a multimodal AI reasoning model to generate structured reports detailing authenticity status, origin verdicts, and confidence levels. The technology is intended for high-stakes applications including legal evidence review, deepfake investigations, journalism, academic integrity, and platform content moderation.

Background: Digital Forensics and AI

← Back to the news feed